Unlocking the Mechanics of Life: Enzymes as Soft, Programmable Nanobots

Abstract
The catalytic cycle involves internal motions and conformational changes that allow enzymes to specifically bind to substrates, reach the transition state and release the product. Such mechanical interactions and motions are often long ranged so that mutations of residues far from the active site can modulate the enzymatic cycle. In particular, regions that undergo high strain during the cycle give mechanical flexibility to the protein, which is crucial for protein motion. Here we directly probe the connection between strain, flexibility and functionality, and we quantify how distant high-strain residues modulate the catalytic function via long-ranged force transduction. We measure the rheological and catalytic properties of wild-type guanylate kinase and of its mutants with a single amino acid replacement in low-/high-strain regions and in binding/non-binding regions. The rheological response of the protein to an applied oscillating force fits a continuum model of a viscoelastic material whose mechanical properties are significantly affected by mutations in high-strain regions, as opposed to mutations in control regions. Furthermore, catalytic activity assays show that mutations in high-strain or binding regions tend to reduce activity, whereas mutations in low-strain, non-binding regions are neutral. These findings suggest that enzymes act as viscoelastic catalytic machines with sequence-encoded mechanical specifications.

Living cells are bustling with molecular machines that constantly process energy, matter, and information. Among these machines, proteins stand out, with enzymes being the most notable. These catalytic entities dramatically accelerate essential metabolic reactions by many orders of magnitude, facilitating the very process that sustain life. While it has long been acknowledged that enzymes undergo movements during their catalytic cycles, measuring and predicting these internal motions and forces has proven extremely challenging.

This study addresses that challenge. It is the result of an international collaboration, led by Professor Tsvi Tlusty from the Department of Physics at UNIST and Professor Elisha Moses from the Physics Department at the Weizmann Institute of Science, Israel.

Figure 1. Schematic representation illustrating the experimental analysis method discovering the highly deformable regions of enzymes, which function like shock observers in mechanical systems. l Image Credit: Weizmann Institute of Science, Israel

The collaboration integrated artificial intelligence (AI) models with molecular dynamics simulations to predict the internal dynamics of enzymes, along side an innovative “nano-rheology” technique to measure these dynamics with unprecedented accuracy. The computational and experimental results culminated in the development of a new viscoelastic model of enzymes, elucidating the intertwined effects of elastic forces arising from stretching or twisting molecular bonds and friction forces (viscosity) associated with bond breaking and reforming.

“Finally, this novel physical model can explain how subtle, nanoscale motions and forces within enzymes impact their biological functions. It allows us to perceive proteins as soft robots or programmable active matter,” said Professor Tlusty.

The findings of this research have been published in Nature Physics on March 28, 2025.

Journal Reference
Eyal Weinreb, John M. McBride, Marta Siek, et al., “Enzymes as Viscoelastic Catalytic Machines,” Nature Physics, (2025).